A new institutional approach to pro-poor agricultural development:Lessons from Asia

James Roumasset *

Dept. of Economics, University of Hawaıi-Manoa, 2424 Maile Way, 542 Saunders Hall, Honolulu, HI 96822, United States

Seiji Naya has been one of only two instructors of a course unique to the University of Hawaii called, ‘‘The Economics ofCooperation.’’ He was an early observer of deliberation councils and other institutions whereby the investment coordinationproblem was partly solved by extramarket cooperation in the East Asian ‘‘miracle’’ countries (Lee & Naya, 1988). Theseinsights contributed to a University of Hawaii–sponsored volume, The Economics of Cooperation (1992), that anticipated the‘‘market-friendly’’ interventionism of the East Asian Miracle (1993) by a full year. A central theme of this special issue—therole of government as facilitator—was also advocated by the State of Hawaii when Dr. Naya served as director of the HawaiiDepartment for Business, Economic Development and Tourism. Reconsidering government as a facilitator, and not as areplacement for markets, is one of the primary contributions of the New Institutional Economics. In this paper, I develop thisperspective in another context: that of agricultural development.

Agricultural growth has long been characterized as stimulating pro-poor economic development, due to its linkages tofood prices, labor demand and growth of the modern sector. Balisacan (2004, 2007) allows that agricultural development iscapable of reducing poverty by 3–4% points for every 1% of economic growth, given the presence of appropriateinfrastructure, trade policies, and institutions supportive of specialization and exchange. That these high elasticities ofpoverty reduction have not been realized in practice is presumably because of misguided policies. Subsidies have acceleratedrent seeking and stagnated growth. Even investments in research and infrastructure have not lived up to their advocates’promises, due to poor incentives and institutional design.

Journal of Asian Economics 19 (2008) 378–388

A R T I C L E I N F O

JEL classification:

Q12

Q18

Q15

Q16

L14

Keywords:

Agricultural policy

Transaction costs

Contracts

Specialization

A B S T R A C T

Dr. Naya has been one of only two instructors of a course unique to the University of

Hawaii called, ‘‘The Economics of Cooperation.’’ He was an early observer of deliberation

councils and other institutions whereby the investment coordination problem was partly

solved by extra-market cooperation in the East Asian ‘‘miracle’’ countries. These insights

contributed to The Economics of Cooperation (1992), which featured the role of government

as facilitator—a theme of Hawaii State Development Planning when Dr. Naya served on the

Governor’s cabinet.

Reconsidering government as a facilitator, and not as a replacement for markets, is one

of the primary contributions of the New Institutional Economics (NIE). This paper extends

and applies the facilitation perspective to the problem of agricultural development. Policy

failures are detailed and sourced to the fallacy of misplaced exogeneity. In contrast, the

method of fundamental explanation, inherent in NIE, acts as a corrective to misguided

interventionism that has prevented pro-poor rural development from taking place.

� 2008 Elsevier Inc. All rights reserved.

* Tel.: + 1 808 956 7496; fax: +1 808 956 4347.

E-mail address: jimr@hawaii.edu.

Contents lists available at ScienceDirect

Journal of Asian Economics

1049-0078/$ – see front matter � 2008 Elsevier Inc. All rights reserved.

doi:10.1016/j.asieco.2008.09.009

In what follows, I provide some examples of policy failures and show how faulty reasoning was partly to blame. I thenprovide an overview of the New Institutional Economics and discuss in various contexts how it acts as a corrective to theinterventionism that has prevented pro-poor rural development from taking place.

1. Policy failures

Agricultural development is the intellectual backwater of economics. Even after the economic developmentinterventionists were driven into retreat during the ‘‘neoliberal’’ 1980s, the economics of agricultural developmentremained largely anti-market. How did this come to pass?

The old interventionism follows the legacy of A.C. Pigou. Markets in rural areas of developing countries areunderdeveloped: they either fail to exist or they exist with widespread externalities. In either case, market failure ispervasive and governments must adjust economic incentives with taxes, subsidies and regulations, including outright bansof some economic activities. Demsetz (1969) christened this form of economic reasoning the Nirvana Fallacy, becausePigouvian reasoning implicitly compared a straw man version of voluntary cooperation with a perfect government.

As economists became increasingly aware of the logical pit of Pigouvian thought and witnessed widespread governmentfailures in both design and implementation of government policies, intellectual support for the old interventionism waned.However, it was quickly replaced by the new interventionism advocated by Stiglitz and followers.

The most general interventionist doctrine is based on the Greenwald and Stiglitz (1986) theorem according to which acompetitive equilibrium is not constrained-Pareto optimal, i.e., it is not on the feasible utility frontier, whose limits aredetermined by feasible government actions as well as technology, factor endowments and consumer preferences. Thistheoretical result is interpreted to mean that government can always find a coercive intervention to increase economicefficiency over that achieved by voluntary contracting and competitive markets. In this ‘‘New Information Economics,’’market failures are not limited to the usual cases of externalities, public goods and nonconvexities, but are far morepervasive, including failures due to moral hazard, adverse selection or other information problems (Stiglitz, 1993).

Similarly, de Janvry, Fafchamps, and Sadoulet (1991), while acknowledging the role that transaction costs play in ruralorganization, nonetheless conclude that ‘‘indirect sources of market failure need to be eliminated’’ including access to creditand insurance markets. Both de Janvry et al. (1991) and de Janvry and Sadoulet (2000) have been misconstrued to mean thatgovernment should intervene in such markets with mandates and subsidies (see e.g., Weber et al., 2002). Some investmentsin agriculture, notably in agricultural research, are prematurely rejected in this view as mere ‘‘technofix.’’

These propositions bear a strong family resemblance to Nirvana Fallacy, however. The equilibrium concept in question isa straw man in two important respects. First, it does not admit multilateral voluntary contracting. Second, it does not admitprivate governance of moral hazard and other information problems, e.g., as described in Jensen (2000). Even if theGreenwald–Stiglitz theorem were generalized to allow for multiple distortions and even if some pervasive efficiency-improving interventions were found, the results would still suffer from blackboard economics.1

1.1. Large farm inefficiency and land-to-the-tiller reform

As an illustration of both forms of interventionism, consider the long-held belief among agricultural economists that theinverse relationship between farm size and yield per hectare is evidence of the inefficiency of large commercial farms. Theold interventionists simply asserted that there was labor market dualism whereby commercial farms paid a higherinstitutional wage than the peasant or subsistence sector. Accordingly, they concluded that land-to-the-tiller reform wouldimprove resource allocation and boost agricultural productivity.2 Clearly, however, they assumed inefficiency in order toconclude there was room for increased efficiency; this is hardly a viable general argument.

The new interventionists provide a more sophisticated explanation of the inverse relationship but draw the same land-to-the-tiller conclusion. Inasmuch as small farms rely on family labor, they are said to economize on the transaction costs ofhiring labor on which commercial farms depend. Accordingly, hired labor is characterized as inefficient (Otsuka, 2002).World Bank (2003) notes that these ‘‘labor market imperfections result in the productive superiority of family farms.’’ Usingthe ICRISAT village data, Frisvold (1994) found that family labor is indeed more productive than hired labor, even beforededucting the costs of supervision. Similarly, Hayami (2003) finds that, while plantation agriculture was an efficientinstitution for the exploitation of Western colonies in Asia, family farms have more recently ‘‘proved to be equally or moreefficient producers of tropical export crops using the family labor of low supervision costs, relative to plantations based onhired labor.’’

However, these studies fail to account for why labor is hired, for which tasks and for the incomplete substitutability ofhired and family labor. They also fail to account entirely for the role of land quality in crop choice and intensity of cultivation.It is not surprising, therefore, that one can find contradictory empirical results. Indeed Benjamin (1992) found that hiredlabor cannot be said to be significantly more or less productive than family labor. This may simply be because there are both

1 Note that ‘‘blackboard economics’’ should not be taken as a general condemnation of rigor, but rather of equilibrium concepts that abstract from real

world institutions, which internalize spillovers and mitigate information problems.2 See for example, Berry and Cline (1979).

J. Roumasset / Journal of Asian Economics 19 (2008) 378–388 379

gains and losses involved. For example, hired labor facilitates specialization. On the prototypical farm in which bothfamily and hired labor are employed, rational choice implies that there will be a nonrandom division of tasks betweenfamily and hired labor and that, at the margin, the difference in their productivities will be equal to the difference inopportunity costs.

Moreover, commercial farms enjoy economies of scale in marketing their products (Reardon, Timmer, Barrett, &Berdegu’e, 2003). The interventionists conclude that this is not a real advantage, however, because small farmers can simplyform cooperatives and exploit the same gains.3 This is a remarkable inconsistency. Small farms are said to be more efficientbecause they avoid the additional contracting costs associated with hired labor. But the proposition that small farms cansimply overcome diseconomies through contracts blithely ignores the requisite contracting costs. Indeed, farmercooperatives are notorious for broken agreements and favoritism, both of which undermine the sustainability of groupcontracts.

1.2. The alleged inefficiency of share tenancy

Relatedly, substantial interventionist ink has been spilled asserting the inefficiency of another institution, i.e., sharetenancy. The old interventionist view was based on the so-called Marshallian model, which was perfect Pigouvianism,albeit before Pigou.4 According to Marshall’s famous footnote, the rational tenant equates his marginal opportunitycost of labor with only his share of the marginal product. This conclusion has been used to justify the other primaryplank of land reform—that is, the banning of share tenancy. Cheung (1969) debunked this view, observing that theMarshallian model could hardly be an equilibrium contractual solution inasmuch as the landlord and tenant couldrenegotiate the share and amount of output or inputs that must be provided or used, thereby making both partiesbetter off.

Stiglitz (1974) proposed a principal agency model wherein sharecropping is viewed as a pairwise-efficient means ofincentivizing labor, relative to wage contracts, without the cost of risk-bearing that would be imposed under rent contracts.He thus resurrected Marshallian inefficiency and the proposition that share tenancy should be outlawed. Indeed Stiglitz(1993, 2002) has often used the institution of share tenancy to exemplify how economic organization can be in equilibriumbut massively inefficient, asserting that a landlord’s output share of 50% would have the same disincentive effects as a 50%income tax. The model has had a long and successful run in agricultural development circles. Hayami and Otsuka (1993)concluded that the risk aversion vs. moral hazard model indeed ‘‘justifies the existence of share tenancy in the theoreticallymost consistent manner’’ and econometric studies (e.g., Shaban, 1987) have concluded that the model is empirically sound.

As is the case with the literature on the inefficiency of large farms and hired labor, however, this conclusion is premature.First, the canonical model does not imply, as originally claimed (Stiglitz, 1974), that the optimal share, b, varies positivelywith the tenant’s degree of risk aversion. Risk aversion also blunts the tenant’s incentive to shirk. Second, the model isincapable of explaining the empirical distributions of tenant shares, which cluster around 50%, with a smaller cluster ofaround 2/3.5 However, the larger problem is that the theory fails to recognize the nature of share tenancy, a typically long-term contractual arrangement for bringing management together with land and that facilitates the tenant’s learning-by-doing about production decisions (Eswaran & Kotwal, 1986; Murrel, 1983; Reid, 1976; Roumasset, 1995). Share tenantsthemselves hire substantial amounts of labor, especially for the more arduous and routine tasks. On the other hand, share

contracting, as distinct from share tenancy, is a popular labor contract for specific tasks. Indeed, share tenants often hirecasual workers on a share basis to do harvesting, weeding and transplanting.

These rationales for land reform fail to acknowledge the complexity of economic cooperation. The principle ofcomparative advantage implies that different characteristics of land and landowners will call for different intensities andcomposition of inputs and organizational forms with unlimited differences in architecture. Judging the relative efficiency ofdifferent organizational forms commits the most fundamental fallacy in economics—judging performance withoutunderstanding the nature and causes of the phenomenon of interest.6 Prescribing policy reforms based on the premise thatpoliticians, bureaucrats and academics can socially engineer institutions superior to those shaped, tested and improved inthe crucible of evolution is a recipe for government failure.

For example, land reform in the Philippines outlawed share tenancy. As a result, land reform beneficiaries hiredpermanent workers who were paid a fixed amount for the season. Hayami and Otsuka (1993) concluded that this has been aninferior substitute for share tenancy. Another Philippine example concerns the failure to properly base landlordcompensation on quality. By basing compensation on the principle that 25% of yield is a fair rent, reform confiscates valuefrom owners of good and average farms but actually overrewards owners of poor quality land (Roumasset & James, 1979). Asa result, friends and relatives of poor quality landowners submit bogus claims that they have been working the land astenants so that the landlord receives more than the land is worth (and landownership remains in the family).

3 See for example, the review of literature and discussion in World Bank (2003).4 It is easy to show that Marshallian underemployment is readily cured by a Pigouvian labor subsidy.5 Deweaver and Roumasset (2002) show that for parameters representative of the Philippine case, the Stiglitz model predicts that the optimal tenant’s

share is U-shaped in tenant’s risk aversion and never falls below 80%.6 In Coasean terms, this is known as blackboard economics.

J. Roumasset / Journal of Asian Economics 19 (2008) 378–388380

1.3. Middlemen: credit and output markets

According to the old interventionism, usury laws are needed to control allegedly exploitative moneylenders (whooften are ethnic Chinese). However, usury laws curtail the amount of loanable funds, cause excess demand and increasethe share of available funds going to the nonpoor. This led to an explosion of directed credit programs beginning inthe 1970s, whereby rural banks were given loanable funds on concessionary terms on the condition that they wouldlend to targeted clients. However, funds were inadvertently diverted to the nonpoor (Meyer & Nagarajan, 2000), and thecredit subsidies indirectly penalized the informal credit sector (Roumasset, 1986). Beneficiaries of subsidized creditcorrectly perceived that the subsidies were part of a system of political patronage and that they were not obligated torepay the loans. Rather than abandon the directed credit approach, however, many governments attempted to patchthem up with new tranches of funding and loan guarantees that, in turn, only increased moral hazard. This is band-aid

economics.The new interventionism only increased support for such programs. Advocates commonly cite Stiglitz and Weiss (1981)

as showing that competitive credit markets are constrained-Pareto inefficient.7 As already discussed, however, these arestraw man arguments in the sense that they fail to account for private governance and voluntary multilateral contracting.

1.4. Uncertainty

The old interventionism asserts that low-income farmers are especially risk averse, that the new technologies offered tothem are highly risky and that they, therefore, underinvest in high-yielding varieties and other components of therecommended package of practices. Thus uncertainty becomes a cover for a panoply of government interventions, from cropinsurance to fertilizer subsidies. The new interventionism augments this case by asserting that moral hazard and adverseselection inevitably cause private insurance markets to fail.

The case for subsidizing crop insurance and allegedly risk-increasing inputs does not hold up to either logical orempirical scrutiny, however. Traditional analysis has equated risk aversion with variance aversion. However, if thethreshold income is high, low-income farmers may be variance preferring. In common parlance, this is known asdesperation.

Empirically, it turns out that modern inputs thought to increase risk often decrease risk or exhibit a U-shaped relationshipbetween risk and the amount of the input applied. In a study of Philippine farmers, it was shown that risk typically decreasesin fertilizer either up to or almost up to the expected profit-maximizing quantity (Roumasset, 1976). Moreover, cropinsurance distorts farmer behavior, causing farmers to neglect available strategies of risk reduction such as pest control(Roumasset, 1979).

The efficiency case for parastatals in developing countries is that a substantial fraction of the population is poor and riskaverse, that markets are incomplete and that insurance fails to compensate for missing state-contingent and futures marketsdue to moral hazard and adverse selection (e.g., Just, 1988). The proper conclusion from these observations is that there may

be some form of government intervention that can result in a welfare improvement from the competitive equilibrium. Bysome leap of imagination, it has often been assumed, however, that the failure of the competitive equilibrium to achieve afirst-best optimum implies that there are benefits from price stabilization.8 For example, in their response, Just (1988) statedthat ‘‘attempts to quantify the net efficiency benefits of institutional attempts to reduce risk suggest that they are usuallysmall and possibly negative.’’9 This, in turn, is interpreted to mean that the theoretical benefits are substantial but may not berealized in practice (Timmer, 1989). Again, these propositions turn out to be wrong in theory (Wright & Williams, 1990) andfallacious in practice (Roumasset, 2003a).

Agricultural parastatals are charged variously with providing low and stable prices to consumers, sufficiently highand stable prices to producers, promoting agricultural modernization, insuring food security and reducing poverty.These goals are not only fiscally irresponsible but are in conflict. Not only does pursuit of some of the goals prevent theachievement of others, but actual programs thwart the development of the private marketing sector; artificially inflateconsumer prices relative to producer prices and motivate corruption. Calculations for the Philippines show thatinterventions in the rice market alone caused excess burden of more than US$ 1 billion in 1 year (David, 2003;Roumasset, 2002).

Parastatals are also illustrative of black hole economics (Roumasset, 2003b). Like drug wars, parastatals increase anddestabilize consumer prices, i.e., make the problem worse. Like prohibition, this induces more government effort, whichfurther worsens the problem, thus setting a vicious circle in motion. In principle, there is no limit to the resources that can besucked into the fiscal black hole.

In summary, interventionist economics is bankrupt, primarily due to a problem of misplaced exogeneity and the NirvanaFallacy. In the next section, I offer an alternative to the Pigouvian blackboard economist and the postmodern informationeconomics.

7 See also Bardhan and Udry (1999).8 The typical scheme envisioned involves a buffer stock administered to maintain domestic prices between set ceiling and floor prices.9 Quoted in Timmer (1989).

J. Roumasset / Journal of Asian Economics 19 (2008) 378–388 381

2. The new institutional economics of agricultural organization10

The alternative to misplaced exogeneity involves characterizing the true nature and seeking the fundamental causes ofbehavior and organization (Barzel, 1989; Coase, 1988). In a cross-section of farms, for example, which type of land isallocated in large parcels, to which economic actors, and why? How has the composition between family and hired laborchanged and why? Under what conditions do landlords choose to contract with tenants to manage their land?

The central decisionmaking model of development microeconomics is the farm-household model. A simple version isdepicted in Fig. 1, which shows the household labor supply schedule of a representative farm-household and three possiblelabor demand schedules, depending on (quality-adjusted) farm size.

For D1, the family exports its excess labor and the relevant shadow price of labor is ws, the ‘‘selling wage’’ after deductingjourney to work and other necessary expenses from the nominal wage. For D3, the farm-household imports hired labor and theshadow wage is wh, the hiring wage after including the employer’s agency cost, which includes recruiting and supervision costsand the residual costs of labor shirking (see discussion above). If labor demand intersects household supply in the intermediaterange between wh and ws, the shadow wage rate is given by the household’s marginal opportunity cost of labor, SL.

11

Accordingly, the rational farm-household can be said to be maximizing shadow profits, based on the shadow wage schedule:

w ¼ ws; L< L1

wh; L> L2

SL; L1 < L< L2

The profit maximization problem of the farm is only quasi-separable from the household utility maximization problem,inasmuch as the labor supply schedule is not independent of farm income.

Similarly, the farm-household produces the shadow profit-maximizing quantity of the agricultural commodity, wherethe shadow price is bounded by the buying price and the selling price, and coincident with the household demand schedulein between. Again there is a limited source of nonseparability, inasmuch as household demand is dependent on farm income.The ‘‘wedge model’’ contrasts with the farm-household model of Lau, Yotopoulos, Chou, and Lin (1981) and Ahn, Singh, andSquire (1981) wherein household consumption is determined recursively, based on the profit-maximizing behavior of thefarm. Nonetheless, a recursive algorithm can be employed to solve the wedge model, albeit by guessing householdconsumption and iterating until the guessed consumption level is consistent with both the household utility function andshadow profit-maximizing farm income.

However, the wedge model begs the question regarding determination of the unit transaction-cost wedge. That isprovided by agency theory. Fig. 2 illustrates agency theory in the context of alternative labor contracts. Piece rates arecommonly used in situations where the product of labor is easily observable, for example, sizing and sharpening the cane

Fig. 1. Quasi-separability of farm labor demand and farm-household supply.

10 A previous version of this section appeared in ‘‘Rural institutions, agricultural development, and pro-poor economic growth,’’ Asian Journal of Agriculture

and Development, 1(1), 61–82 (June 2004).11 For further details of this model, see Roumasset (1981). A similar model and circumscribed comparative statics are provided in de Janvry et al. (1991)

and Sadoulet and de Janvry (1995). Roumasset (1979) extends the model to include behavior under uncertainty.

J. Roumasset / Journal of Asian Economics 19 (2008) 378–388382

stalks prior to planting, and the planting of stalks at uniform spacing. These tasks are tantamount to intermediate productsdelivered to the farm operator, who pays according to quantity. This institution economizes on minimum agency cost, i.e.,the minimum sum of supervision and shirking costs. For tasks that are not amenable to ex post inspection, supervision isused to concurrently monitor the labor activity in question and workers are paid according to the time spent on an activity,not its result. The four panels illustrate the comparative statics proposition that if tasks are sufficiently easy to monitorthrough ex post inspection, then the corresponding agency cost at optimal monitoring will be lower than the agency cost ofwage contracts. The opposite is true for tasks that are difficult to monitor. For each task, the unit transaction cost is given bythe least of the two minimum agency costs (MAC) for the task in question.

The wedge model can be used to explain behavior of the farm-household—the basic building block for theories ofagricultural development. The agency cost model can be used for explaining rural institutions. Both are essential forunderstanding the consequences of contemplated policy reforms.

The new institutional economics of agricultural organization (Roumasset, 1974, 1978) also recognizes that differentlevels of analysis may be appropriate for the analysis of different problems. Models that recognize transaction costs such asthe two above are classified as second best.12 When the subject of inquiry is the terms of agricultural organization—e.g.,tenant and harvesters share of production—the first-best model, which abstracts from transaction costs, has been found to beappropriate. In first-best analysis, the terms of contracts are set such that factors receive their marginal products, just as ifthere were competitive markets.13 Third-best analysis or political economy allows for multilateral opportunism in thepursuit of favorable government treatment by special interests (Dixit, 1996).

Fig. 2. Specialization of contracts by task.

12 Note that while both models accommodate transaction costs, the first regards them as being exogenous while the second determines unit transaction

costs endogenously.13 This is the implicit theoretical underpinning of Hayami and Kikuchi’s (1982) study of rural institutions in the Philippines and Indonesia. Sufficient

assumptions and a theoretical demonstration of market and contract equivalence are provided in Roumasset (1979).

J. Roumasset / Journal of Asian Economics 19 (2008) 378–388 383

2.1. Land, labor and the nature of the farm

Consider the evolution of hired labor. In the Marxist view, the new rice and wheat technologies that swept throughAsia in the 1970s disenfranchised the peasantry and led to falling wages and increased unemployment. In the inducedinnovation view (Binswanger & Ruttan, 1978; Ruttan, 2003), the causation was just the reverse. Population pressure onlimited land resources drove down wages thereby inducing land-saving technological change. In effect, this allowed‘‘biological capital’’ (modern varieties and chemical inputs) and labor to substitute for land. The increased demand forlabor had a positive effect on wages, just not enough to offset the effect of population pressure (Hayami & Kikuchi,1982).

The induced-technological-change explanation just described is a first-best argument. However, not only did labor perhectare increase, its composition changed dramatically. In the 10 years following the adoption of the new rice varieties in thePhilippines, hired labor in weeding, for example, increased from less than 20% of total labor to more than 80% (Roumasset &Smith, 1981). Fig. 3 illustrates the use of the wedge model to explain this dramatic institutional change.

The graph represents a typical farm-household in the province of Laguna and shows how four factors combined toincrease hired labor dramatically. First, and most importantly, the intensification of production—ultimately caused byincreasing land scarcity and accommodated by the new rice technology—increased the demand for labor per hectare. This isillustrated by the shift in the demand curve to the right. Second, increased farmer incomes resulted in increased schooling offarm children. This combined with the increased specialization among farm workers lowered the amount of farm-householdlabor per hectare. These higher opportunity costs and lower substitutability for skilled labor are illustrated by the shift in thelabor supply curve to the left. Third, the market wage went down (from Wm0 to Wm1) as population growth, includinginmigration, increased by more than enough to supply the increased labor demand. Fourth, the transaction-cost wedgebetween the market wage and the gross hiring wage decreased due to the advent of labor contractors and other newinstitutions of labor contracting (Roumasset & Uy, 1980). The third and fourth factors are illustrated by a downward shift inthe gross hiring wage (from Wh0 to Wh1).

As hired labor increased, a menu of agricultural contracts emerged for incentivizing labor in different tasks. We havealready discussed in Fig. 2, which shows how agency theory can be used to explain the tendency for piece rate contracts to bechosen when the task amounts to delivering an observable intermediate product. Statistical analysis of sugarcane contractsin the Philippines confirms this tendency (Roumasset & Uy, 1980). For example, cane stalks are prepared for planting(uniformly sized and sharpened) and laid out for inspection. The farm operator simply inspects them for quality anduniformity. Next the stalks are planted and the operator inspects for proper height and spacing.

Gama or ilani, as practiced in the Philippines, is an institutional arrangement whereby the worker contracts to weed andharvest a specified parcel for typically one-sixth of the rice harvested for that parcel; ceblokan, practiced in Indonesia,typically requires transplanting, in addition to harvesting and weeding, for the same one-sixth share (Hayami & Kikuchi,1982; Roumasset, 1978).14 These arrangements were preceded by hunusan in the Philippines and bawon in Indonesia,wherein only harvesting was done for the share of the harvest, typically one-sixth. Before the new institutions of gama andceblokan, the share was sometimes lowered to one-eighth (Roumasset, 1978).

Why did the share settle at one-sixth and the work increase instead of the share simply declining? Hayami (1998)suggested that another function of gama/ceblokan was to provide an explicit selection mechanism for choosing who wouldweed/harvest and to allocate a specific parcel to each group of workers. In addition to selection, this provides improvedincentives over the open hunusan/bawon systems that were open to anyone in the village. Under the old system, a kind of freeriding occurred wherein workers would harvest faster than efficiency warrants in the attempt to harvest more land than

Fig. 3. HYVs and the advent of labor markets.

14 Remarkably, a similar arrangement was documented in The Constitution of Athens almost 3000 years ago. Workers contracted under a sharing

arrangement in ancient Greece were called hectomori or ‘‘sixth partners’’.

J. Roumasset / Journal of Asian Economics 19 (2008) 378–388384

their competing harvesters. Moreover, having workers harvest the same plot that they weeded (and sometimestransplanted) provided additional incentives to weed/transplant with greater care. Thus while first-best principles canexplain either the falling harvesters share or the increased work required, second-best considerations are required tounderstand why one institution was favored over the other.

Fig. 4 provides a second-best efficiency explanation of the institution of share tenancy. The larger the tenant share, theless the agency costs of labor shirking (defined as monitoring cost plus residual shirking costs). On the other hand, the greaterthe tenant share, the greater the tenant’s incentive to overuse or undermaintain land quality. Share tenancy (with a tenant’sshare of roughly one-half) minimizes the agency cost of both sources combined. There is nothing inherently inefficient in thecontract, just explicit recognition of the contracting costs inherent in specialization.

Inasmuch as the tenant is the farm manager, not a worker, it is futile to classify ‘‘forms of tenure’’ as share tenant, lessee orwage worker. Rather, we need to classify organizational forms by which ownership, management and labor are connected. Fig. 5illustrates a taxonomy of firms classified according to degree of specialization. Note that the pure owner–operator and theowner–manager are on opposite sides of the specialization spectrum, even though the conventional taxonomy classifies themboth as owner–operator. The pure owner–operator household does all the management and all the labor: there is no hired labor.The owner–manager hires most of the labor and reserves for himself only those tasks, which are bundled with managerialdiscretion, e.g., fertilization. Share tenancy is characterized by an intermediate amount of specialization whereby the tenantdoes most of the management, all discretionary tasks and some other tasks, e.g., land preparation. Evidence from the Philippinesand Nepal confirms that specialization is driven by intensity of cultivation, which is driven, in turn, by favorable land quality,location and economic environment (Roumasset, 1995). Intensification can also be driven by population pressure, demandgrowth and rising land values. Not only does intensification warrant more specialized agricultural firms, but the organization ofhired labor itself becomes increasingly specialized. This is elaborated in the following section.

2.2. The nature of economic integration: transaction costs and specialization

As farm production intensifies, labor inputs increase, until the last stage wherein capital–labor substitution overcomesinput intensification. Labor contracts are increasingly specialized, eventually with labor contracts made on a task-by-taskbasis. Thus intensification and specialization are co-evolutionary. Fig. 6 helps to resolve the fundamental paradox that totaltransaction costs increase as economic development proceeds (North & Wallis, 1982). Unit costs of transportation andcommunication (unit transaction costs) are falling and improved institutions lower the agency costs (defined as supervisionplus residual shirking costs) per unit of labor hired. However, because more labor is hired and because specializationincreases the number of contracts (even normalized by yield per hectare), transaction expenditures increase.

Fig. 5. A spectrum of agricultural firms.

Fig. 4. An Eclectic theory of share tenancy.

J. Roumasset / Journal of Asian Economics 19 (2008) 378–388 385

2.3. Facilitating specialization in input and output markets

As unit transaction costs fall and incomes rise with economic development, the number of both final and intermediategoods increases as does the number of distinct labor tasks and opportunities for learning-by-doing. Thus economicspecialization and integration are part of the same evolutionary process (Yang, 2003). However, natural market deepening isimpeded by market-distorting interventions including trade restrictions, price interventions, shipping and other regulations,and failure to provide public infrastructure, including quality standards. Parastatals, such as the National Food Authority inthe Philippines, exemplify how government policy can stagnate the natural evolutionary process and thereby stagnate anindustry. Economic integration can be enhanced by removing these policy distortions and by focusing on facilitating actionssuch as agricultural research and the provision of transportation and communication infrastructure.15

By considering institutional choice as endogenous, we can understand two beneficial effects that are often overlooked.First, inasmuch as institutional change is induced by changing factor prices (Ruttan, 1978, 2003)—e.g., falling wages relativeto rents—it allows greater substitution of labor for land, thus partially ameliorating downward pressure on wages. Second, tothe extent that institutional change facilitates specialization and the external economies associated therewith (Yang, 2003),it may actually overcome the original downward pressure on wages (Roumasset, 2008; Roumasset & Van Assche, 2003).Econometric studies showing that hired labor is less productive than family labor fail to account for the specialization goingon and for the fact that the farm operator’s labor is considerably more valuable than the shadow price of hired labor. In otherwords, the inefficiency arguments ignore the principle of comparative advantage.

3. Conclusion

Coase (1988) once remarked that his work has been ‘‘much cited but little used.’’ This is still the case, at least indevelopment economics. For example, the World Bank (2007) persists in searching for market failures, apparently unawarethat this approach has consistently resulted in policy failures as exemplified by land reform policies, credit subsidies,marketing parastatals, and crop insurance.

Writing down a model of climate change that fits the stylized facts of temperature does not prove that global warming iscaused by greenhouse gases. Similarly, writing down inefficiency models that purport to explain the inverse relation

Fig. 6. Evolution of labor contracts.

15 Frequent reference to Philippine agriculture as the least integrated in southeast Asia is misleading. Statistical measures of integration have not been

informed by correct theory. The naive measures used presume that equality of shadow prices across space and across economic agents is the efficient

benchmark. Even more sophisticated theory that equates shadow price differentials with transport costs is correct only where positive transport costs exist.

Moreover, it is misleading to separate space from time. As noted above, optimal trade and transportation of grain calls for exporting from the south

following their peak harvest and importing to Manila preceding the wet season harvest on Luzon. During periods when the efficient transport cost is zero,

shadow prices differentials can be less than transport costs.

J. Roumasset / Journal of Asian Economics 19 (2008) 378–388386

between farm size and yield per hectare do not prove that commercial farms are inefficient. As von Neumann put it, ‘‘Withfour parameters I can fit an elephant, and with five I can make him wiggle his trunk.’’ Just as radiative forcing models must beconsidered alongside alternative models of climate change, so should inefficiency models be considered along with othermodels that yield the same statistical patterns. If statistical tests fail to discriminate between competing theories policyreforms should allow for the possibility of diverse causes. In particular, premature diagnosis of market failure is a recipe forgovernment failure (Besley, 1994).

But statistical analysis is not the only means by which theories can be distinguished. Occum’s Razor counsels prefersimpler and less general theories when added complexities buy us nothing in terms of explanation of observables.Fundamental theories are preferred to ad hoc theories. Explaining the inverse relationship with a model that assumesdualism in the labor market is not a valid method for prescribing policy reform. The new institutional economics provides afundamental method for explaining agricultural organization that does not rest on misplaced exogeneity.

Looking for fundamental causes of institutional change engenders a healthy skepticism of social engineering. Rather thefocus of economic policy reform should be on facilitating economic cooperation through the provision of information, a legalinfrastructure and opportunities for multilateral cooperation. The prerequisites for cooperation will render the time-honored strategy of pushing agricultural development through investments in research and infrastructure even moreeffective, especially if modern principles of public administration are employed (Laffont & Tirole, 1993).

The economics of rural organization with endogenous behavior and organization is in its infancy. There is a promisingbody of theory featuring specialization as the central pillar of economic organization (e.g., Yang, 2003) and a rich tapestry ofrural institutions waiting to be described and explained.

References

Ahn, C. Y., Singh, I., & Squire, L. (1981). A model of an agricultural household in a multi-crop economy: The case of Korea. Review of Economics and Statistics, 63, 520–525.

Balisacan, A. (2004). Averting hunger and food insecurity in Asia. Asian Journal of Agricultural Development, 1(1), 36–55.Balisacan, A. (2007). Local growth and poverty reduction. In A. Balisacan & H. Hill (Eds.), The dynamics of regional development the Philippines in East Asia. Quezon

City and Cheltenham: Ateneo de Manila Press and Edward Elgar.Bardhan, P., & Udry, C. (1999). Development microeconomics. Oxford: Oxford University Press.Barzel, Y. (1989). Economic analysis of property rights. Cambridge: Cambridge University Press.Benjamin, D. (1992). Household composition, labor markets, and labor demand: Testing for separation in agricultural household models. Econometrica, 60, 287–

322.Berry, A., & Cline, W. (1979). Agrarian structure and productivity in developing countries. Baltimore: Johns Hopkins University Press.Besley, T. (1994). How do market failures justify interventions in rural credit markets? World Bank Research Observer, 9, 27–47.Binswanger, H., & Ruttan, V. (1978). Induced innovation. Baltimore: Johns Hopkins University Press.Cheung, S. (1969). Transaction costs, risk aversion, and the choice of contractual arrangements. Journal of Law and Economics, 12, 23–42.Coase, R. (1988). The firm the market and the law. Chicago: University of Chicago Press.David, C. (2003). In A. Balisacan & H. Hill (Eds.), The Philippine economy: Development, policies, and challenges. New York: Oxford University Press. Chapter 6.de Janvry, A., & Sadoulet, E. (2000). Poverty determinants in Latin America. Food Policy, 25, 389–409.de Janvry, A., Fafchamps, M., & Sadoulet, E. (1991). Peasant household behavior with missing markets: Some paradoxes explained. Economic Journal, 101, 1400–

1417.Demsetz, H. (1969). Information and efficiency: Another viewpoint. Journal of Law and Economics, 12, 1–22.Deweaver, M., & Roumasset, J. (2002). Risk aversion as effort incentive: A correction and primafacie test of the canonical theory of share tenancy. Economics

Bulletin, 15, 1–16.Dixit, A. (1996). The making of economic policy: A transaction-cost politics perspective. Cambridge: MIT Press.Eswaran, M., & Kotwal, A. (1986). Access to capital and agrarian production organization. Economic Journal, 96, 482–498.Frisvold, G. (1994). Does supervision matter? Some hypothesis tests using Indian farm-level data. Journal of Development Economics, 43, 217–238.Greenwald, B., & Stiglitz, J. (1986). Externalities with imperfect information. Quarterly Journal of Economics, 101, 229–264.Hayami, Y. (1998). Norms and rationality in the evolution of economic systems: A view from Asian villages. Japanese Economic Review, 49, 36–53.Hayami, Y. (July 2003). Family farms and plantations in tropical development. FASID discussion paper 2003-001.Hayami, Y., & Kikuchi, M. (1982). Asian village at the crossroads: An economic approach to institutional change. Baltimore: Johns Hopkins University Press.Hayami, Y., & Otsuka, K. (1993). The Economics of contract choice: An Agrarian perspective. Oxford: Clarendon Press.Jensen, M. (2000). A theory of the firm: Governance, residual claims, and organizational forms. Cambridge: Harvard University Press.Just, R. E. (1988). Making economic welfare analysis useful in the policy process: Implications of the public choice literature. American Journal of Agricultural

Economics, 70, 448–453.Laffont, J.-J., & Tirole, J. (1993). A theory of incentives in procurement and regulation. Cambridge and London: MIT Press.Lau, L., Yotopoulos, P., Chou, E., & Lin, W.-L. (1981). The microeconomics of distribution. Journal of Policy Modeling, 3, 175–206.Lee, C., & Naya, S. (1988). Trade in east Asian development with comparative reference to southeast Asian experiences. Economic Development and Cultural Change,

36, S123–S152.Meyer, R., & Nagarajan, G. (2000). Rural financial markets in Asia: Policies, paradigms, and performance. Oxford: Oxford University Press for the Asian Development

Bank.Murrel, P. (1983). The economics of sharing: A transaction cost analysis of contractual choice in farming. Bell Journal of Economics, 14, 283–293.North, D. C., & Wallis, J. J. (1982). American government expenditures: A historical perspective. American Economic Review, 72, 336–340.Otsuka, K. (2002). Land markets. In Bretton Woods conference.Reardon, T., Timmer, P., Barrett, C. P., & Berdegu’e, J. (2003). The rise of supermarkets in Africa, Asia, and Latin America. American Journal of Agricultural Economics,

85, 1140–1146.Reid, J. D. (1976). Sharecropping and agricultural uncertainty. Economic Development and Cultural Change, 24, 549–576.Roumasset, J. (1974). Induced institutional change, welfare economics, and the science of public policy. Working Paper No. 46. Economics Dept., University of

California-Davis (as cited in Ruttan, V. W. Induced institutional change. In B. Hans, V. Ruttan, et al. (Eds.). Induced innovation: Technology, institutions, anddevelopment. Baltimore/London: The Johns Hopkins University Press).

Roumasset, J. (1976). Rice and risk: Decision-making among low-income farmers in theory and practice. Amsterdam: North-Holland Publishing.Roumasset, J. (1978). The new institutional economics and agricultural organization. Philippine Economic Journal, 37, 331–348.Roumasset, J. (1979). Risk aversion, agricultural development, and the indirect utility function. In J. Roumasset, J.-M. Boussard, & I. Singh (Eds.), Risk, uncertainty

and agricultural development. Manila, Philippines: SEARCA and Agricultural Development Council.

J. Roumasset / Journal of Asian Economics 19 (2008) 378–388 387

Roumasset, J. (1981). Positive methods of agricultural decision analysis. Development Studies Centre, Australian National University Press.Roumasset, J. (1986). The welfare economics of rural credit and the benefits of land titles. World Bank manuscript.Roumasset, J. (1995). The nature of the agricultural firm. Journal of Economic Behavior and Organization, 26, 171–177.Roumasset, J. (2002). The microeconomics of agricultural development in the Philippines. Working paper # 02-10-2002. University of Hawaii, Department of

Economics.Roumasset, J. (2003a). The role of agriculture in pro-poor economic development: A Philippine perspective, Mimeo.Roumasset, J. (September 2003b). Black hole economics. Hawaii Reporter. <http://www.hawaiireporter.com/file.aspx?Guid=e5519258-c009-449b-99f5-

84c0ea027330>.Roumasset, J. (2004). Rural institutions, agricultural development, and pro-poor economic growth. Asian Journal of Agriculture and Development, 1(1), 61–82 Los

Banos, Philippines: Southeast Asian Regional Center for Graduate Study and Research in Agriculture.Roumasset, J. (2008). Population and agricultural growth. In S. N. Durlauf & L. E. Blume (Eds.), The New Palgrave Dictionary of economics (2nd ed.). Palgrave Macmillan.

The New Palgrave Dictionary of economics online. Palgrave Macmillan.<http://www.dictionaryofeconomics.com/article?id=pde2008_A000065> (10.07.2008),doi:10.1057/9780230226203.1308.

Roumasset, J., & James, W. (1979). Explaining variations in share contracts: Land quality, population pressure and technological change. Australian Journal ofAgricultural Economics, 23, 116–127.

Roumasset, J., & Smith, J. (1981). Population, specialization, and landless workers: Explaining transitions in agricultural organization. Population and DevelopmentReview, 7, 401–419.

Roumasset, J., & Uy, M. (1980). Piece rates, time rates and teams: Explaining patterns in the employment relation. Journal of Economic Behavior and Organization, 1,343–360.

Roumasset, J., & Van Assche, A. (2003). Malthus to solow: A new classical approach to induced innovation. Presented at the Cliometrics Society meetings, June 3–5,2003, Granlibakken Conference Center, Lake Tahoe, California.

Ruttan, V. (1978). Induced institutional change. In H. P. Binswanger & V. Ruttan (Eds.), Induced innovation. Baltimore: Johns Hopkins.Ruttan, V. (2003). Social science knowledge and economic development: An institutional design perspective. Chicago: University of Michigan Press.Sadoulet, E., & de Janvry, A. (1995). quantitative development policy analysis. Baltimore: Johns Hopkins University Press.Shaban, R. (1987). Testing between competing models of share tenancy. Journal of Political Economy, 95, 893–920.Stiglitz, J. (1974). Incentives and risk sharing in sharecropping. Review of Economic Studies, 41, 219–256.Stiglitz, J. (1993). Whither socialism? Cambridge: MIT Press.Stiglitz, J. (2002). Globalization and its discontents. New York: Norton.Stiglitz, J., & Weiss, A. (1981). Credit rationing in markets with imperfect information. American Economic Review, 71, 393–410.Timmer, P. C. (1989). Food price policy: The rationale for government intervention. Food Policy, 14, 17–27.Weber, C., Franklin, D., Diaz, E., Bathnik, D., Zuvekas, C., & Southgate, D. (2002). Rural prosperity in the Latin American and Caribbean region. USAID/LAC White paper.

Washington, DC.World Bank (2003). Land policies for growth and poverty reduction: A World Bank policy research report. New York: World Bank/Oxford University Press.World Bank. (2007). World development report 2008: Agriculture for development. Washington, DC: The World Bank.Wright, B., & Williams, J. (1990). The behavior of markets for storable commodities. Paper presented at the annual conference of the Australian Agricultural Economic

Society.Yang, X. (2003). Development economics: Inframarginal vs. marginal analysis. Oxford/Melbourne: Blackwell.

J. Roumasset / Journal of Asian Economics 19 (2008) 378–388388